A brand defense campaign that shows a low CPA and high conversion rate looks like a clear win — until you realize most of those installs would have happened anyway, organically, for free. Measuring real cannibalization is the only way to know if brand defense spend is protecting revenue or just replacing it.
Bidding on your own brand name in Apple Search Ads feels intuitively defensive: if you don't, a competitor might buy that placement and intercept someone who already searched for you by name. That instinct is reasonable, but it skips a harder question — how many of the installs your brand campaign reports were actually incremental, versus installs you'd have received organically at zero cost either way?
Cannibalization happens when a store visitor installs your app from a paid ad, which you pay for, even though they would have installed it anyway without the ad — through your organic listing, at no cost. Brand defense campaigns are especially prone to this because they target searchers who already know your app by name and are highly likely to convert with or without a paid placement in front of them. A campaign report showing a low CPA and high conversion rate on branded terms can look like a strong result while actually representing near-total cannibalization — the metrics look good because the underlying searcher was always going to install, not because the ad caused anything.
Illustrative proportions — the real split can only be measured for your own app, not assumed from a diagram.
The most direct way to measure real incrementality is to pause your brand defense campaign for one to two weeks and monitor what happens to your total branded-term installs (paid plus organic combined). If total installs on your brand terms drop by roughly the same amount as the paid campaign was reporting, cannibalization was low — the ad was genuinely capturing installs that wouldn't have happened otherwise. If total installs barely move during the pause, that's strong evidence the campaign was mostly replacing organic traffic you'd have gotten anyway. The installs "lost" during the pause approximate your protected installs; the difference between the campaign's reported installs and that protected number approximates cannibalized installs.
Estimated protected (incremental) installs: 60
Estimated cannibalization rate: 85%
A more rigorous, ongoing approach than a single pause test is a structured incrementality analysis — comparing a geographic or user-segment holdout (no brand ads shown) against a treatment group (brand ads running as normal) over a longer window, isolating what the ads actually drove versus what would have happened anyway. This requires more setup than a simple pause test, but it produces an ongoing incrementality estimate rather than a single-point snapshot, which matters because cannibalization rates can shift as your organic ranking strength, competitor bidding behavior, and brand awareness all change over time.
Directionally true, unverified: cannibalization rates vary significantly by app category, brand strength, and competitive pressure on your branded terms — don't assume a published industry average applies directly to your own account without measuring it.
It's worth being explicit about why this problem is so easy to miss without a pause test. Attribution systems generally credit an install to the last ad a user interacted with before installing — so if someone searches your brand name, sees your ad (which they'd likely have tapped through to your organic listing anyway), and installs, the attribution system correctly records that the ad was present, but it has no way of knowing whether that person would have installed without ever seeing the ad at all. This isn't a flaw in Apple's attribution reporting specifically — it's a structural limitation of last-touch attribution in general, and it applies to brand campaigns more than almost any other campaign type precisely because branded searchers already have high purchase (or install) intent walking in the door.
Your existing organic rank on your own brand term matters directly to how much cannibalization risk you're carrying. If you already rank #1 organically for your own app name — which is the common case for an established app with a distinct name — a searcher typing your name in is extremely likely to find and tap your listing with or without a paid unit above it. If your organic rank on your own brand term is weaker (perhaps due to a generic name that collides with other apps or common search terms), a paid brand campaign has a more plausible protective function, since the organic listing alone might not reliably surface first. This is one of the clearest, cheapest diagnostic checks to run before even setting up a pause test: check your own organic rank on your own brand name.
Cannibalization isn't automatically a reason to abandon brand defense entirely. If competitors are actively bidding on your brand name, an unprotected top slot can genuinely be intercepted, and the real comparison isn't "pay for these installs vs. get them free" — it's "pay for these installs vs. risk losing some of them to a competitor's ad." The decision should weigh your measured protected-install rate against the real competitive risk in your category: a low-cannibalization campaign in a highly contested category is a much stronger case for continued spend than a high-cannibalization campaign in a category where no competitor is bidding on your name at all.
"NoteFlow," a note-taking app, runs a two-week pause test on its brand campaign. These are illustrative numbers to show the calculation, not real measured data.
| Period | Total branded installs/week | Paid campaign spend |
|---|---|---|
| Campaign running | 500 | $450 |
| Campaign paused | 430 | $0 |
Protected installs ≈ 70/week. True cost per protected install ≈ $450 / 70 ≈ $6.43 — a very different number than the campaign's reported blended CPA, which likely looked much lower because it was averaged across mostly-cannibalized installs.
Illustrative relationship for demonstration — actual cannibalization depends on your own measured pause-test data, not competitive pressure alone.
If a pause test reveals high cannibalization and low competitive risk, the honest next step isn't just "cut the brand campaign" — it's redirecting that freed budget toward campaigns with a much clearer incrementality story: discovery keywords where you don't already rank highly organically, category terms where you're competing for genuinely new searchers, or a Custom Product Page-backed campaign targeting a specific underserved intent. Budget cut from a high-cannibalization brand campaign without a plan for where it goes tends to just get reabsorbed into general marketing spend with no clear attribution story of its own — the value of this whole exercise comes from redeploying the freed spend somewhere it can be measured as genuinely incremental, not just from the act of cutting itself.
Cannibalization isn't a fact you establish once and file away — organic rank strength, brand awareness, and competitive bidding pressure on your name all shift over quarters, sometimes meaningfully. An app that measured low cannibalization a year ago because a new competitor hadn't yet started bidding on its brand term could have a very different risk profile today. A practical habit is a lightweight quarterly check: re-run a short pause test (even just a few days, cross-checked against the same weekday pattern as your original longer test) and re-check your organic brand-term rank, rather than assuming the original measurement still holds indefinitely. This is a small recurring cost compared to running an entire brand campaign on outdated assumptions about its own efficiency.
A brand defense campaign's reported metrics can look great while representing mostly cannibalized installs you'd have gotten organically for free. Measure real incrementality with a pause test (or a more rigorous holdout analysis), calculate true cost per protected install rather than blended CPA, and weigh that against genuine competitive risk on your brand terms before deciding how much to spend.